Psychosis frequently occurs as a symptom of neurodegenerative disease, but this phenomenon is understudied and widely misunderstood. It can often be the first sign of disease, which has direct relevance to clinicians' ability to identiy the earliest brain networks affected in disease, and thus the ability to predict the underlying neuropathology. Psychosis also places a heavy burden on medical and social services, and exacerbates existing strain on dementia patients, their families, and their social networks. The overarching goal of this project is to perform secondary analysis of two large existing datasets to identify the clinical phenomenology, neuroanatomy, and neuropathology underlying different types of delusions and hallucinations in three diseases of aging: Alzheimer's disease (AD), behavioral variant frontotemporal dementia (bvFTD) and dementia with Lewy bodies (DLB). The two research databases to be analyzed were collected at the UCSF Memory and Aging Center through funding provided by the NIH National Institute of Aging. First, in an autopsy-confirmed pathology cohort of 172 cases, we will determine how distinct types of delusions and hallucinations are able to predict the different underlying neuropathological diagnoses. Second, in a larger cohort of 540 clinically diagnosed patients with AD, bvFTD and DLB, existing structural MRI data will be quantified using voxel-based morphometry in order to identify brain regions that are associated with distinct psychotic presentations. Finally, functional connectivity analysis will be conducted on resting state fMRI data from 210 subjects to identify more subtle links between specific types of psychosis and dysfunction in specific brain networks. First, we hypothesize that the basis for psychosis is the dysfunction of a core network involved in appraisal of situational cues that produces faulty inferences about reality, and that this dysfunction occurs as a result of rostromedial prefrontal cortex damage and disconnection, and is common across neurodegenerative diseases. Second, there is a distinct sensory input system that is uniquely affected in each disease, which dictates the subtype and content of the psychosis. In bvFTD, faulty visceral sensory input, resulting from insular damage and salience network dysfunction, leads to bizarre and body-focused psychosis. In AD, faulty episodic and autobiographical memory input, resulting from hippocampal damage and memory network dysfunction, leads to non-bizarre paranoid delusions. In DLB, faulty sensory visual input, resulting from occipitotemporal damage and visual network dysfunction, leads to misidentification delusions and visual hallucinations. Thus, more precise clinical characterization of psychotic symptoms can yield critical information about the focal brain network affected, which in turn will lead to earlier, more accurate diagnosis for patients with these diseases.